PDA

View Full Version : R134a <-> liquid heat exchanger design







DaBit
07-05-2002, 02:57 PM
This topic is a followup of <a href="http://www.refrigeration-engineer.com/chat/showthread.php?s=&threadid=706">this</a> topic, to stay ontopic.

For the newcomers: I have built an R134a based water chiller which cools a water/methanol mixture, which in turn should cool the microprocessor in my PC. I am currently able to reach -16 &deg;C under load, but I am pretty sure my current evaporator is not performing too well. It's leaky too. For more details see the previous topic.

So, I need a new heat heat exchanger. First I was after a plate heat exchanger, but I got a few answers from the manufacturers which all say that my system is too light for a plate heat exchanger.

Heat exchanger design parameters:

- Qmax=200W @ -20 &deg;C
- Qnom=150W @ -20 &deg;C
- Refrigerant used: R134a (~4kg/hour)
- Coolant used: 30/70 vol.% methanol/water mixture.
- Coolant flow: 600-1000L/hour, netto.

For this new heat exchanger I was thinking of a coaxial heat exchanger design using a 3/8" copper tube (evaporator) inside a 22mm (approx. 7/8") copper pipe.

I also want to equip the inner pipe with copper fins over about 2/3 of it's length, to increase the contact area with the coolant. Why not over the full length? Well, I need some space to allow the coolant to distribute and mix. Fin layout will be something like this:


ffffffff ffffffff ffffffff
====================================
ffffffff ffffffff ffffffff


The 3/8" inner copper tube will give me a decent refrigerant speed (at the end of the evaporator around 750ft/min)

The 22mm outer copper tube carries the coolant and it guarantees a decent coolant speed, although I prefer it to be higher. But the next smaller size pipe is 15mm, and this imposes a too high flow restriction.

The piping itself will be bent into a circular/spiral like construction to save space. Refrigerant will enter on top, suction gas will leave at the bottom, so refrigerant and oil flow is supported by normal gravity.

Questions:

- Is this design OK, or can it be better?
- How long should I make the coaxial pipe construction? I think 3 .. 3.5 meters (10 feet) will be sufficient.

zolar1
07-05-2002, 03:14 PM
Many years ago, people used to use an old compressor to make a home made vacuum pump.

Once the system has been dehydrated enough, you shouldn't need to pull a vacuum for such a long period of time.
On something that small, an hour or 2 might be appropriate.

If you open the system to make design changes, make sure you plug off the open lines to the compressor and replace the dryer (preferably with a bigger one).

The oil in an R134a system is hydroscopic (it sucks up water like a sponge) and doesn't want to give it back. Heating the compressor to 400 deg F for a few hours while under high vacuum was suggested by another site I went to.

You could probably do better with R22 sice it's compatable with most oils (See ICOR International), can get colder, and is cheaper.

You could braze a 'Tee' onto your water inlet (inside the H.E.) and braze 2-45 deg elbows on the open parts at such an angle so as to make the incoming water swirl

DaBit
07-05-2002, 03:54 PM
Many years ago, people used to use an old compressor to make a home made vacuum pump.

Once the system has been dehydrated enough, you shouldn't need to pull a vacuum for such a long period of time.
On something that small, an hour or 2 might be appropriate.


Is the vacuum pulled by a normal working LBP R134a compressor enough to dehydrate the system? And does such a compressor keep running for a few hours without breaking down (provided that there is a decent airflow over the thing to keep it cool)?


You could probably do better with R22 sice it's compatable with most oils (See ICOR International), can get colder, and is cheaper.

No, R134a is going to be the refrigerant. I am unable to obtain any other refrigerant than R134a and BBQ grade propane (poor man's R290). Here in The Netherlands it is even illegal to use R134a without a license.
Besides that, R22 pressures are too high for the compressor and condenser.
But I don't believe that a high pressure refrigerant would perform better. The R134a compressor just runs at a lower suction pressure, and therefore the compressor has a larger volumetric displacement per Watt of motor power. This is also how R600a compressors work. R600a suction pressure is very low which is compensated with a large displacement compressor.
Once you go to really low temperatures this might change since then pressures are almost vacuum, and the compressor would not displace much gas per cycle anymore.

When comparing compressors for the same temperature range, with the same motor power, but only a different refrigerant, you will see approximately similar numbers for capacity.



You could braze a 'Tee' onto your water inlet (inside the H.E.) and braze 2-45 deg elbows on the open parts at such an angle so as to make the incoming water swirl


I think coolant flow is already quite turbulent inside the tube. The inner walls of the tube cause friction which alters flow. Otherwise, a better way to create more turbulence is to bend the end of the fins a bit as to create rotational motion in the coolant flow.

Question: would it be wise to switch over to a small MOP type TXV instead of a captube? Or is 4kg/hour refrigerant flow not enough for even the smallest TXV? Will it cause valve hunting?

zolar1
07-05-2002, 04:21 PM
If I made a different but similar heat exchanger for you, would you be willing to try it? If so, let me know.

DaBit
07-05-2002, 04:40 PM
If I think it will perform better than the coaxial design I presented here, sure!

zolar1
07-05-2002, 05:03 PM
It will take me a little time to (pardon the pun) 'Engineer' one for you.

Once I get it done, I'll post here again so you can email me your mailing address.

zolar1
08-05-2002, 05:27 AM
I think I can improve on your design a little.

I need to know what size tubing you need to adapt to.

You mentioned 1/4" for the refrigerant, but what about the water inlet and outlet?

How cold do you want the water to be?
Does the physical size matter or does it have to be small?

The larger the size of the container holding the water mix, the greater the load can be placed on the whole system (intermittently)

You mentioned the PVC piping expanded at a different rate than the copper. I think you can still use the PVC, just use a bigger diameter and longer length. A baffle would be in order between the water inlet and outlet, much like a steam generator.

Still thinking about the design changes needed, but need more info please.
Can you use 1/4" inlet and 5/16" or 3/8" outlet for the refrigerant?

DaBit
08-05-2002, 09:22 AM
Originally posted by zolar1
[B]I think I can improve on your design a little.

I need to know what size tubing you need to adapt to.
You mentioned 1/4" for the refrigerant, but what about the water inlet and outlet?


Well, if possible I want to switch to an expansion valve which has a 1/4" liquid connection. The 1/2" connection coming out of the valve is not useable as is anyway since 1/2" evaporator pipes are too thick to ensure enough refrigerant flow. 3/8" is a more suitable thickness for the evaporator piping.



How cold do you want the water to be?
Does the physical size matter or does it have to be small?


My goal is a liquid temperature of -20 &deg;C (-4F), and the compressor is able to handle that.

Physical size does matter. The enclosure for the entire cooling system will measure 65x40x35 centimeters.



The larger the size of the container holding the water mix, the greater the load can be placed on the whole system (intermittently)


This is _not_ the case. The load varies depending on how busy the thing is, but it does not really feature a 'duty cycle'

Therefore, I want the amount of coolant in the heatex to be quite small, so the temperature drops rapidly.



You mentioned the PVC piping expanded at a different rate than the copper. I think you can still use the PVC, just use a bigger diameter and longer length. A baffle would be in order between the water inlet and outlet, much like a steam generator.


I want to get rid of the PVC anyway. It is too hard to seal and PVC is not very good in handling low temps.



Still thinking about the design changes needed, but need more info please.
Can you use 1/4" inl%t and 5/16" or 3/8" outlet for the refrigerant?

The compressor has a 3/8" suction line connection. And let's keep the liquid line 1/4".

zolar1
08-05-2002, 12:39 PM
OK..as I see it, you have for the refrigerant lines:

1/4" inlet to the H.E.
3/8" outlet from the H.E.

What about the inlet and outlet sizes for the water lines?

DaBit
08-05-2002, 12:44 PM
Nope, 3/8" inlet to the HE, adapted from 1/2" expansion valve output. The expansion valve wants a 1/4" liquid line as input.

Water connections are 12-15mm O.D. copper pipes.

Gary
08-05-2002, 11:47 PM
In the interests of improving communication, what you are calling a heat exchanger is what we would call an evaporator. In our trade, when we talk about heat exchangers, we are generally referring to suction/liquid line heat exchangers.

Your new evaporator design looks fine to me, but keep in mind that system design is outside my area of expertise. On the other hand, once it is up and running, I can tell you if you did it right, and what needs to be done to make it work. I don't calculate and predict, I measure and analyze.

I would go with the TEV, and the dual heat exchangers we discussed in the other thread. It should work very well. The low starting torque of your compressor may be a problem, but hard start kits are easy to install.

Dan
09-05-2002, 02:10 AM
what happens when you refrigerate a light bulb? Does the wattage change? Does the light dim? How is a light bulb different from a computer chip? What is similar?

zolar1
09-05-2002, 07:18 AM
Is the light bulb fluorescent or incandesant?

Light bulb is one heat generator, a cpu is many heat generators (transistors on = more heat)

Both use power.

Gary, don't you mean Chiller instead of Heat exchanger?

By definition, a heat exchanger is a device that exchanges (transfers) heat from one source to another. So, a chiller would be a heat exchanger too.

Gary
09-05-2002, 11:24 AM
The condenser also transfers heat, but we call it a condenser.

This entire system could be called a chiller, but when I say evaporator, you know exactly what I mean, right?

DaBit
09-05-2002, 12:03 PM
Your new evaporator design looks fine to me, but keep in mind that system design is outside my area of expertise. On the other hand, once it is up and running, I can tell you if you did it right, and what needs to be done to make it work. I don't calculate and predict, I measure and analyze.

Before you build you should calculate and predict, otherwise you don't know what to build.
After you are finished building, you should measure and analyze.



I would go with the TEV, and the dual heat exchangers we discussed in the other thread. It should work very well.


I already planned to do so. I am trying to arrange a Danfoss T2 TXV with the smallest orifice. And or course; the dual suction gas<->liquid heat exchanger system also looks OK.

I have one question left: Why not place the TXV bulb after both heat exchangers? Enough superheat is enough superheat, right?




The low starting torque of your compressor may be a problem, but hard start kits are easy to install.

All I see mentioned is that, by adding a start capacitor and replacing the starting relay, I can get an increased torque. I have no idea whether or not this is enough torque. The compressor is and stays designed for captube use.

Gary
09-05-2002, 12:58 PM
I have one question left: Why not place the TXV bulb after both heat exchangers? Enough superheat is enough superheat, right?

The idea is to flood the evaporator with liquid, but not flood the compressor.

The first heat exchanger warms the bulb, providing more liquid to the evaporator, ensuring that it is thoroughly flooded, but still keeping the TXV within it's control range.

The second heat exchanger adds heat to the returning refrigerant to boil off any remaining droplets, thus ensuring that no liquid reaches the compressor.

Both heat exchangers drop the temperature of the liquid entering the TXV, which increases the efficiency of the evaporator. It is a very effective setup.


All I see mentioned is that, by adding a start capacitor and replacing the starting relay, I can get an increased torque. I have no idea whether or not this is enough torque. The compressor is and stays designed for captube use.

The difference between a high torque compressor and a low torque compressor is the start capacitor.

Gary
09-05-2002, 01:03 PM
Before you build you should calculate and predict, otherwise you don't know what to build.

Exactly. That's what design engineers do.


After you are finished building, you should measure and analyze.

And that's what I do.

DaBit
09-05-2002, 08:21 PM
The second heat exchanger adds heat to the returning refrigerant to boil off any remaining droplets, thus ensuring that no liquid reaches the compressor.

Both heat exchangers drop the temperature of the liquid entering the TXV, which increases the efficiency of the evaporator. It is a very effective setup.

OK. I already planned to do the double heat exchanger setup.

Question: does brazing the liquid line and suction line together provide enough heat exchange, or am I better off running the liquid line inside the suction line (also a coaxial construction)?


The difference between a high torque compressor and a low torque compressor is the start capacitor.

Now this is what I wanted to hear! This removes all the doubt about what to do. I can go for a start capacitor without having a bad feeling about it. Thanks.


Exactly. That's what design engineers do.

People who have analyzed and measured enough are often more capable of designing a system than people who only have their books and simulation programs ;)

zolar1
10-05-2002, 08:22 AM
In carpentry, what is the transfer method? Why do people sometimes use it? How would it relate to refrigeration and design?

zolar1
10-05-2002, 08:36 AM
A condenser transfers heat, but it's name was derived from it's function - to condense.

Isn't a heat exchanger's name derived from it's function too?